Will epigenetics tell us anything useful about Neandertals?

A new paper last week by David Gokhman and colleagues described the pattern of methylation in the high-coverage ancient genomes from Denisova Cave.

Here is the abstract of the study:

Ancient DNA sequencing has recently provided high-coverage archaic human genomes. However, the evolution of epigenetic regulation along the human lineage remains largely unexplored. We reconstructed the full DNA methylation maps of the Neandertal and the Denisovan by harnessing the natural degradation processes of methylated and unmethylated cytosines. Comparing these ancient methylation maps to those of present-day humans, we identified ~2000 differentially methylated regions (DMRs). Particularly, we found substantial methylation changes in the HOXD cluster that may explain anatomical differences between archaic and present-day humans. Additionally, we found that DMRs are significantly more likely to be associated with diseases. This study provides insight into the epigenetic landscape of our closest evolutionary relatives and opens a window to explore the epigenomes of extinct species.

Methylation is a chemical modification to DNA, mostly involving cytosine residues that occur immediately next to guanine (so-called CpG sites). The level of methylation of CpG sites in the DNA in normal cells is very high, upward of 80% or more, but promoter regions of genes tend to be less methylated, and methylation of these DNA region is inversely related to gene expression in cells.

Gokhman and colleagues were able to infer the degree of methylation in these ancient genomes by examining the distinctive pattern of damage to the DNA. Cytosine residues in ancient sequences are often deaminated, which changes them to uracil. Until this pattern was recognized, it was a major source of errors in the interpretation of ancient sequences; now the pattern of deamination has become useful as a way of recognizing genuine ancient sequences as opposed to contaminating modern sequence. The probability of deamination varies with methylation of the cytosine, and this gives a way of interpreting methylation of the original DNA.

Methylation patterns are potentially interesting as indicators of the activity of genes in ancient cells. These patterns vary among tissues in the human body, as methylation is one way that pluripotent stem cells become differentiated into functional types. Hence, bone cells (including ancient bone cells) are different from other tissues in the body in their patterns of methylation. Moreover, because methylation of these somatic cells is not inherited in the germline, there is a lot of “noise” present in the signal of methylation in cells. Some methylation may be purely ideosyncratic, while some is strongly functional and directly determined by gene sequences.

Overall, the study finds that the methylation of ancient DNA in these bones is the same as that present in living people. But there are a good number of locations where the methylation in the ancient bones is significantly higher or lower than found in some living people, and these are potentially interesting areas of further investigation. The most newsworthy is an increase in the methylation of the HOXD9 promoter and HOXD10 gene in the ancient genomes, relative to a lower level of methylation in bone of living people. The study suggests that this methylation change may be a correlate of the limb morphology of Neandertals, in particular with their broader and relatively larger joint surfaces, curved bone shafts and shorter limbs.

I’m skeptical. As Gokhman and colleagues recognize, we have a long way before we will understand the variation of methylation patterns within human populations. They make a start with a very limited number of samples of bones of contemporary people:

A difference in methylation between an archaic and a present-day human does not necessarily imply a fixed difference between the human groups. This difference can stem from variability within a population; or from the comparison of close, yet not identical, cell types (osteoblasts versus whole bones). Hence, we compared the archaic methylation to 37 bones samples, taken from osteoblasts and whole bones (12–14). We sought reliable DMRs, in which the archaic methylation significantly differs from that in modern bones. These samples were measured with 27K arrays and provided information for ~5% of DMRs. In most DMRs, archaic methylation was significantly different from the 37 bones, therefore classified as reliable (FDR < 0.01, z-test, Tables S2, S3).

It’s a start, but far from being sufficient to show that these ancient DNA methylation patterns are very different from living people. “Differentially methylated” is a very interesting concept, because methylation is at the borderline of continuous versus discrete variation. We will likely be able to find living people who share a similar pattern of methylation in these gene regions, and thereby investigate whether their phenotypes approach those of Neandertals. They may not be identical to the ancient genomes, but there is no question that the biology of living people overlaps with Neandertals in many of these phenotypic measures. We do have people with broad joint surfaces, relatively short limbs and robust hands and feet today. The exact morphology is not the same, but the ranges of variation overlap – enough so to make a similar causal pathway a good hypothesis.

We should be able to investigate the variation within ancient populations in a similar way. The variation among Neandertals has been obscured by bad biological stereotyping. Sure, some of the European Neandertals have these aspects of limb morphology, but the West Asian Neandertals extend the overall variation substantially toward longer limbs with less curvature. The pattern of variation in Neandertals is complex, in other words.

It doesn’t help that the high-coverage genome from Denisova is geographically far from any Neandertal skeletal remains that actually have limbs. And the Denisovans have no skeletal remains with limbs at all. I’ll be more comfortable with this kind of work when more actual phenotypic evidence is at hand!


Gokhman D, Lavi E, Prüfer K, Fraga MF, Riancho JA, Kelso J, Pääbo S, Meshorer E, Carmel L. 2014. Reconstructing the DNA Methylation Maps of the Neandertal and the Denisovan. Science (in press) doi:10.1126/science.1250368